An Experiment: Cupping Temperatures

Today I ran a quick experiment for the morning’s cupping. Willy Hansen had kindly sent us from coffee from Solberg & Hansen, and I used their Kenya Karamikui for our test. We’ve been using the Uber boiler to cup for some time now, and it great to remove another variable. I had been wondering, however, how important water temperature was.

The experiment was fairly simple. We cupped the same coffees at 9 different temperatures. Starting at 87C up to 95C. The temperature here was the temperature at the exit of the spout. We tried to let them cool a little to even the odds and I asked everybody to come and cup. I explained that all the bowls had the same coffee in, one factor was different and I asked everyone to pick a favourite bowl.

The results weren’t what I was expecting, if I am honest. I expect massive differences in the cups – especially at the extremes. This wasn’t the case. All cups were tasty, but the lower (brewing) temp cups didn’t do quite as well cold. Temperature of the liquid when cupped definitely affected things – with everyone preferring (unsurprisingly) the bowls that had had longer to cool. Once temperatures started to even out, the hotter brews started to shine through.

If we’d done espresso at these temperatures I am sure the differences would have been a little more obvious. I spent bits of today wondering why it wasn’t so extreme. I concocted a very quick test:

Hat-tip to Mr Truby for the repurposed Hario inspiration

I put a probe into the cups and logged two different brewing temperatures – the two extremes of that day’s cuppings, 87C and 95C.

Click to embiggen

Looking at the chart there are a few interesting things.  Both brews lose an equal amount of heat – about 9 degrees.  Cooling with the crust on appears more uniform.  You can clearly see the point where I break the bowl, and then skim/clean.

Ambient temperature in the room was quite high – 26.5C (roaster is still running here, it’s been a long day….) and the period of time for the experiment was about 25 minutes per bowl.  Ambient was cooler earlier in the day, by a few degrees anyway.

If you had all 9 brews on the chart then I guess things would look pretty close, but if I am honest then I have to say that I am surprised that it didn’t impact flavour, mouthfeel and acidity more.  It possibly didn’t help that the coffee was so good to start with!

I’d like to do a few more tests at future cuppings, but for now I thought I would throw the info out there to see what people think.


  1. Was the cupping spoon used rinsed in hot or cold water and would that make a difference as it would draw some of the heat out of the bowl if it was cold to start with.

    I’m not sure it would have much of an influence but it is another variable.

    Good post and certainly food for thought. Interesting graph results

  2. Are we talking about the cupping or the logging experiment. In the logging experiment the spoon was clean and cold both times.

    In the cupping the spoon was preheated, and rinsed in hot water.

  3. Very interesting experiment. One possible explanation to offer. Based on my current understanding, the higher the water temperature the higher the rate of extraction. The possible reason then for the lower brew temp cups not faring so well, is that whilst still tasty they may have been marginally under extracted compared to the higher temp cups. Out of curiosity, did you measure the brews once cooled?

  4. Interesting- I am surprised you didn’t find a larger difference in quality/taste.
    Here is a half-baked thought… What if there is an ideal temp x time x pressure (and maybe coarseness or some other variable?)for coffee extraction (that might differ per coffee).
    So say 95C x 4 minutes x 1 bar pressure for a cupping yields a well extracted cup (gives you a score of 380). If you increase pressure (as in espresso) and want to keep a similar score- you would get time=380/(95×9) or about 27 seconds which would make sense… Changing the temperature by a few degrees in a cupping only changes the score marginally (90Cx4 minutes x 1bar = 360). Perhaps further experimentation would reveal what score results in a bad cup- perhaps you can push it to 330 or 420 but not 300, or 500…

  5. Very interesting James, I’d be curious as to the MoJo results on the cups across the range. An excellent coffee like Karamikui may remain tasty at different extractions but the numbers would still be there

  6. different compounds have different solubility points, so i’m guessing some coffees will have traits that one can pick up more easily with more thorough extraction (higher temps). these attributes may not be the positive qualities of acidity and mouthfeel, but defects, bag, mustiness etc which is where cupping has been used in a ‘normal’ (commodity) coffee environment: to pick up faults. cupping convention that stems from assessing ‘quality’ on other grounds :)

  7. Have run a few similar experiments here. Always interesting, never what you’d expect. At one point, we ran a cupping of 4 different temps on a coffee that was occasionally showing up as “excessively pulpy” to see if brew temp could be causing the fluctuation in perception. Found that the high extraction temp (98C) and the low (90C) were the closest to ferment flavor, while the two in the middle were not-so-pulpy. Also ran a similar probe in two cups at different temps, however they reached a point of equilibrium for me (in an admittedly cooler environmental temp) at/around 75C. Fascinating stuff.

  8. That is interesting. I am now regretting not doing more testing at higher temps. I’d like to do 94-99C (or as close as possible) to see what happens…

    Thanks for sharing!

  9. It’s about quality my dear James.
    Great beans and roasted to perfection.

    ciao willy

  10. These experiments are always inspiring… Do you have the moisture content of the green and roasted coffee? Do you think roast level would make a difference in the outcome? It wold be interesting to to try a natural processed coffees (sugar content could be a little higher). Man, this experiment is overwhelming me with ideas… Thanks! <3

Submit a comment